We just released our first dataset of 10M highly curated TikTok videos. Would love feedback!
If you are training or finetuning video models and need curated datasets, feel free to reachout!
i would love a link
The-data-company/TikTok-10M on hf
I tried loading it using datasets.load_dataset and it worked after grabbing the CSV link directly from the GitHub repo. Make sure to use streaming=True if memory is an issue.
I load it with load_dataset using the raw CSV URL, then switch on streaming so I don’t run out of memory while scanning or sampling videos.